source/Dow Jones Industrial Average Prediction with Media Channel Info-with Social Info.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dow Jones Industrial Average (DJIA) Prediction\n",
    "## 1. Given: data from Aug 1, 2016 to Nov 30, 2017 with DJIA, 30 Companies price & social media likes\n",
    "    12 months Date(T), Company Name, Stock Price, Facebook likes, Facebook disklikes, twitter likes, DJIA\n",
    "    Facebook likes = ('haha')+('like')+('love')+('wow');\n",
    "    Facebook bads = ('sad')+('angry');\n",
    "## Task:\n",
    "###    Given data from Aug 1, 2016 to Nov 30, 2017\n",
    "###    Split data to\n",
    "    1. Training set from Aug 1, 2016 to Jul 31, 2017\n",
    "    2. Validation set from Aug 1, 2017 to Sep 17, 2017\n",
    "    \n",
    "## 2. To Predict T+1 and T+30 of DJIA  in next 1 months (Sep 18, 2017 to Oct 18, 2017).\n",
    "    \n",
    "## Data Sources:\n",
    "### a. Facebook Likes, Dislike\n",
    "### b. Twitter Likes\n",
    "### c. Yahoo Finance 30 companies stock price + DJIA\n",
    "\n",
    "for 30 companies\n",
    "format=>No. company: facebook id\n",
    "1. 3M: 3m\n",
    "2. American Express: AmericanExpressUS\n",
    "3. Apple: AppleInc.HD\n",
    "4. Boeing: Boeing\n",
    "5. Caterpillar: caterpillar\n",
    "6. Chevron: Chevron\n",
    "7. Cisco Systems: Cisco\n",
    "8. Coca-Cola: CocaColaUnitedStates\n",
    "9. DowDuPont: dupontco\n",
    "10. ExxonMobil: ExxonMobil\n",
    "11. General Electric: GE\n",
    "12. Goldman Sachs:goldmansachs\n",
    "13. IBM: ibm\n",
    "14. Intel: Intel\n",
    "15. Johnson & Johnson: jnj\n",
    "16. JPMorgan Chase: jpmorganchase\n",
    "17. McDonald's: McDonaldsUS\n",
    "18. Merck: MerckInvents\n",
    "19. Microsoft: Microsoft\n",
    "20. Nike: Nike, nikesportswear\n",
    "21. Pfizer: Pfizer\n",
    "22. Procter & Gamble: proctergamble\n",
    "23. The Home Depot: homedepot\n",
    "24. Travelers: travelers\n",
    "25. United Technologies: unitedtechnologiescorp\n",
    "26. UnitedHealth Group: unitedhealthgroup\n",
    "27. Verizon: Verizon\n",
    "28. Visa: VisaUnitedStates\n",
    "29. Walmart: Walmart\n",
    "30. Walt Disney: Disney\n",
    "\n",
    "Query target string: 3m, AmericanExpressUS, AppleInc.HD, Boeing, caterpillar, Chevron, Cisco, CocaColaUnitedStates, dupontco, ExxonMobil, GE, goldmansachs, ibm, Intel, jnj, jpmorganchase, McDonaldsUS, MerckInvents, Microsoft, Nike, nikesportswear, Pfizer, proctergamble, homedepot, travelers, unitedtechnologiescorp, unitedhealthgroup, Verizon, VisaUnitedStates, Walmart, Disney\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import datetime as dt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read training and test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dataset = pd.read_csv('integrated_data.csv', encoding = \"utf-8\", parse_dates=[\"date\"],\n",
    "date_parser=lambda x: pd.to_datetime(x, format=\"%Y%m%d %H:%M:%S\"))\n",
    "# convert date to int\n",
    "dataset['date']=dataset['date'].dt.strftime('%Y%m%d')\n",
    "dataset.set_index(['date', 'company'], inplace=True)\n",
    "dataset = dataset.sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "?dataset.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Select the features for the model.\n",
    "### stock open, close, high, low price, trade volumne, facebook likes, facebook dislike, twitte likes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>open</th>\n",
       "      <th>high</th>\n",
       "      <th>low</th>\n",
       "      <th>close</th>\n",
       "      <th>volume</th>\n",
       "      <th>fblikes</th>\n",
       "      <th>fbdislikes</th>\n",
       "      <th>twittelikes</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th>company</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">20160801</th>\n",
       "      <th>3M</th>\n",
       "      <td>177.080002</td>\n",
       "      <td>178.539993</td>\n",
       "      <td>177.080002</td>\n",
       "      <td>178.289993</td>\n",
       "      <td>1598400</td>\n",
       "      <td>246</td>\n",
       "      <td>0</td>\n",
       "      <td>186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>American Express</th>\n",
       "      <td>64.489998</td>\n",
       "      <td>64.739998</td>\n",
       "      <td>63.959999</td>\n",
       "      <td>64.129997</td>\n",
       "      <td>2913900</td>\n",
       "      <td>115</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Apple</th>\n",
       "      <td>104.410004</td>\n",
       "      <td>106.150002</td>\n",
       "      <td>104.410004</td>\n",
       "      <td>106.050003</td>\n",
       "      <td>38167900</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Boeing</th>\n",
       "      <td>133.210007</td>\n",
       "      <td>134.229996</td>\n",
       "      <td>132.699997</td>\n",
       "      <td>133.210007</td>\n",
       "      <td>3067300</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Caterpillar</th>\n",
       "      <td>82.760002</td>\n",
       "      <td>82.879997</td>\n",
       "      <td>81.529999</td>\n",
       "      <td>81.949997</td>\n",
       "      <td>3377100</td>\n",
       "      <td>208</td>\n",
       "      <td>2</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 open        high         low       close  \\\n",
       "date     company                                                            \n",
       "20160801 3M                177.080002  178.539993  177.080002  178.289993   \n",
       "         American Express   64.489998   64.739998   63.959999   64.129997   \n",
       "         Apple             104.410004  106.150002  104.410004  106.050003   \n",
       "         Boeing            133.210007  134.229996  132.699997  133.210007   \n",
       "         Caterpillar        82.760002   82.879997   81.529999   81.949997   \n",
       "\n",
       "                             volume  fblikes  fbdislikes  twittelikes  \n",
       "date     company                                                       \n",
       "20160801 3M                 1598400      246           0          186  \n",
       "         American Express   2913900      115           0            0  \n",
       "         Apple             38167900        0           0            0  \n",
       "         Boeing             3067300        0           0          105  \n",
       "         Caterpillar        3377100      208           2           29  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Features selection\n",
    "X = dataset.loc[:, ['open','high','low', 'close','volume', 'fblikes', 'fbdislikes','twittelikes']]\n",
    "\n",
    "# Get DJIA close index price\n",
    "djia = dataset.loc[(dataset.index.get_level_values('company') == 'DJIA')]\n",
    "djia = djia.loc[:,['close']]\n",
    "\n",
    "X.head(5)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th>company</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>20160801</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18404.50977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160802</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18313.76953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160803</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18355.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160804</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18352.05078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160805</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18543.52930</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        close\n",
       "date     company             \n",
       "20160801 DJIA     18404.50977\n",
       "20160802 DJIA     18313.76953\n",
       "20160803 DJIA     18355.00000\n",
       "20160804 DJIA     18352.05078\n",
       "20160805 DJIA     18543.52930"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "djia.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Prepare T+1 and T+30 DJIA data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>close</th>\n",
       "      <th>T+1</th>\n",
       "      <th>T+30</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th>company</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>20160801</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18404.50977</td>\n",
       "      <td>18313.76953</td>\n",
       "      <td>18066.75000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160802</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18313.76953</td>\n",
       "      <td>18355.00000</td>\n",
       "      <td>18034.76953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160803</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18355.00000</td>\n",
       "      <td>18352.05078</td>\n",
       "      <td>18212.48047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160804</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18352.05078</td>\n",
       "      <td>18543.52930</td>\n",
       "      <td>18123.80078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20160805</th>\n",
       "      <th>DJIA</th>\n",
       "      <td>18543.52930</td>\n",
       "      <td>18529.28906</td>\n",
       "      <td>18120.16992</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        close          T+1         T+30\n",
       "date     company                                       \n",
       "20160801 DJIA     18404.50977  18313.76953  18066.75000\n",
       "20160802 DJIA     18313.76953  18355.00000  18034.76953\n",
       "20160803 DJIA     18355.00000  18352.05078  18212.48047\n",
       "20160804 DJIA     18352.05078  18543.52930  18123.80078\n",
       "20160805 DJIA     18543.52930  18529.28906  18120.16992"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get y value for close(T+1) and close(T+30)\n",
    "_tmp = djia\n",
    "djia['T+1'] = _tmp['close'].shift(-1)\n",
    "djia['T+30'] = _tmp['close'].shift(-30)\n",
    "djia.head(5)   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# shift training data 30 day with 30 stocks + 1 DJIA\n",
    "# Get Total data set from Aug 1, 2016 to Oct 18, 2017\n",
    "total_row = len(X)\n",
    "data_range = total_row -(30*31)\n",
    "_tmpX = X.head(data_range)\n",
    "_tmpX.head(5)\n",
    "_tmpX.tail(5)\n",
    "X = _tmpX"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# drop NaN rows.\n",
    "djia = djia.dropna()\n",
    "# X = X.dropna()\n",
    "#djia.tail(5)\n",
    "# X.tail(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Transform the T+1 and T+30 DJIA from index to 1(growth) or 0 (not growth) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['0', '0'],\n",
       " ['1', '0'],\n",
       " ['0', '0'],\n",
       " ['1', '0'],\n",
       " ['0', '0'],\n",
       " ['1', '0'],\n",
       " ['0', '0'],\n",
       " ['1', '0'],\n",
       " ['0', '0'],\n",
       " ['1', '0']]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# Calculate the trend on T+1 and T+30\n",
    "y = []\n",
    "\n",
    "for index, row in djia.iterrows():\n",
    "    _tmp = []\n",
    "    if (row['T+1']-row['close'] > 0):\n",
    "        _tmp.append('1')\n",
    "    else:\n",
    "        _tmp.append('0')\n",
    "    if (row['T+30']-row['close'] > 0):\n",
    "        _tmp.append('1')\n",
    "    else:\n",
    "        _tmp.append('0')   \n",
    "    y.append(_tmp)\n",
    "    \n",
    "#djia.head(5)\n",
    "y[0:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Remove Company name from dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9548, 9)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x = X.reset_index()\n",
    "x = x.as_matrix()\n",
    "x = np.delete(x, np.s_[1:2], axis=1) #Remove company\n",
    "x.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Standardization¶"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  0.00000000e+00,   6.68093396e-03,   6.71494901e-03, ...,\n",
       "          1.48420767e-03,   0.00000000e+00,   3.11782356e-03],\n",
       "       [  0.00000000e+00,   1.79932989e-03,   1.79914844e-03, ...,\n",
       "          6.93836918e-04,   0.00000000e+00,   0.00000000e+00],\n",
       "       [  0.00000000e+00,   3.53015551e-03,   3.58792990e-03, ...,\n",
       "          0.00000000e+00,   0.00000000e+00,   0.00000000e+00],\n",
       "       ..., \n",
       "       [  1.00000000e+00,   3.66933256e-03,   3.66784397e-03, ...,\n",
       "          0.00000000e+00,   0.00000000e+00,   8.04599628e-04],\n",
       "       [  1.00000000e+00,   2.73845024e-03,   2.73349638e-03, ...,\n",
       "          0.00000000e+00,   0.00000000e+00,   1.57567427e-03],\n",
       "       [  1.00000000e+00,   3.26784342e-03,   3.25704191e-03, ...,\n",
       "          1.15267429e-01,   1.33502861e-02,   0.00000000e+00]])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "mmc = MinMaxScaler(feature_range=(0,1)) # default (0, 1) for sigmoid function\n",
    "mmc.fit(x)\n",
    "x_std = mmc.transform(x)\n",
    "x_std"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. Reshape Data to 3D (time, companies+DJIA, and 9 features) after Standarlization\n",
    "### Split data to training set (92.8%) and testing set (7.2%)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Reshape data to time slot, company+djia + 9 features\n",
    "data = x_std.reshape(308, 31, 9)\n",
    "data\n",
    "y = np.array(y)\n",
    "\n",
    "# # data.shape\n",
    "_X_train, X_test_std, _y_train, y_test = train_test_split(data, y, test_size = 0.072, shuffle = False, stratify = None)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Split traing set to training data(88.5% of training set) and validation data(11.5% of training set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "# Create training and validation sets\n",
    "X_train_std, X_validate_std, y_train, y_validate = train_test_split(_X_train, _y_train, test_size = 0.115, shuffle = False, stratify = None)\n",
    "# X_validate_std[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(252, 31, 9)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_std.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 8. Build model: 3 Layers (128-64-32) LSTM for this prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "### LSTM Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import StratifiedShuffleSplit\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from keras.utils import np_utils\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten, TimeDistributed, LSTM, GRU\n",
    "from keras.callbacks import EarlyStopping, ReduceLROnPlateau, ModelCheckpoint, Callback\n",
    "\n",
    "def create_LSTM_model(optimizer='adam', kernel_init='normal', recurrent_init='normal', bias_init='zero', batch_size=1):\n",
    "    model = Sequential()\n",
    "    #input method: batch_input_shape=(batch, timesteps/lags, features)\n",
    "    #  or input_shape=(timesteps/lags, features) , model.fit(X, y, epochs=n, batch_size=b, verbose=2)\n",
    "    model.add(LSTM(128, batch_input_shape=(batch_size, 31, 9), return_sequences=True, activation='sigmoid', recurrent_activation='hard_sigmoid', use_bias=True, \n",
    "                   kernel_initializer=kernel_init, recurrent_initializer=recurrent_init, bias_initializer=bias_init,\n",
    "                   unit_forget_bias=True, kernel_regularizer=None, recurrent_regularizer=None, bias_regularizer=None, \n",
    "                   activity_regularizer=None, kernel_constraint=None, recurrent_constraint=None, bias_constraint=None, \n",
    "                   dropout=0.0, recurrent_dropout=0.0, implementation=1, return_state=False, \n",
    "                   go_backwards=False, stateful=False, unroll=False))\n",
    "    model.add(Dropout(0.2))\n",
    "    model.add(LSTM(64, return_sequences=True, activation='sigmoid', recurrent_activation='hard_sigmoid', use_bias=True, \n",
    "                   kernel_initializer=kernel_init, recurrent_initializer=recurrent_init, bias_initializer=bias_init, unit_forget_bias=True, \n",
    "                   kernel_regularizer=None, recurrent_regularizer=None, bias_regularizer=None, activity_regularizer=None, \n",
    "                   kernel_constraint=None, recurrent_constraint=None, bias_constraint=None, dropout=0.0, recurrent_dropout=0.0, \n",
    "                   implementation=1, return_state=False, go_backwards=False, stateful=False, unroll=False))\n",
    "    model.add(Dropout(0.2))\n",
    "    # The last layer of LSTM, return_sequences = False\n",
    "    model.add(LSTM(32, return_state=False, activation='sigmoid', recurrent_activation='hard_sigmoid', use_bias=True, \n",
    "                   kernel_initializer=kernel_init, recurrent_initializer=recurrent_init, bias_initializer=bias_init, unit_forget_bias=True, \n",
    "                   kernel_regularizer=None, recurrent_regularizer=None, bias_regularizer=None, activity_regularizer=None, \n",
    "                   kernel_constraint=None, recurrent_constraint=None, bias_constraint=None, dropout=0.0, recurrent_dropout=0.0, \n",
    "                   implementation=1, return_sequences=False, go_backwards=False, stateful=False, unroll=False))\n",
    "    model.add(Dropout(0.2))    \n",
    "    model.add(Dense(2, activation='sigmoid', kernel_initializer=kernel_init))\n",
    "    model.compile(optimizer=optimizer, loss='binary_crossentropy', metrics=['accuracy'])\n",
    "    return model\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(252, 31, 9)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_3D_train_std = X_train_std\n",
    "X_3D_train_std.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 252 samples, validate on 33 samples\n",
      "Epoch 1/20\n",
      "252/252 [==============================] - 13s - loss: 0.6574 - acc: 0.5833 - val_loss: 0.4570 - val_acc: 0.8485\n",
      "Epoch 2/20\n",
      "252/252 [==============================] - 11s - loss: 0.6088 - acc: 0.6528 - val_loss: 0.4122 - val_acc: 0.8485\n",
      "Epoch 3/20\n",
      "252/252 [==============================] - 12s - loss: 0.6163 - acc: 0.6508 - val_loss: 0.4120 - val_acc: 0.8485\n",
      "Epoch 4/20\n",
      "252/252 [==============================] - 11s - loss: 0.6192 - acc: 0.6528 - val_loss: 0.4130 - val_acc: 0.8485\n",
      "Epoch 5/20\n",
      "251/252 [============================>.] - ETA: 0s - loss: 0.6238 - acc: 0.6494\n",
      "Epoch 00004: reducing learning rate to 0.00020000000949949026.\n",
      "252/252 [==============================] - 13s - loss: 0.6227 - acc: 0.6508 - val_loss: 0.4165 - val_acc: 0.8485\n",
      "Epoch 6/20\n",
      "251/252 [============================>.] - ETA: 0s - loss: 0.6151 - acc: 0.6614\n",
      "Epoch 00005: reducing learning rate to 0.0001.\n",
      "252/252 [==============================] - 13s - loss: 0.6144 - acc: 0.6607 - val_loss: 0.4231 - val_acc: 0.8485\n",
      "Epoch 00005: early stopping\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1ce0a6dde10>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Parameters\n",
    "from keras.optimizers import Adam\n",
    "# adam = Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=0.0) #default\n",
    "adam = Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=0.01, decay=0.0)\n",
    "optimizers = adam\n",
    "kernel_init = 'glorot_uniform'\n",
    "recurrent_init = 'glorot_uniform'\n",
    "bias_init ='zero'\n",
    "epochs = 20\n",
    "batches = 1\n",
    "\n",
    "model_LSTM = create_LSTM_model(optimizer=optimizers, batch_size=batches, kernel_init=kernel_init, \n",
    "                 recurrent_init=recurrent_init, bias_init=bias_init)\n",
    "model_LSTM.fit(X_3D_train_std, y_train,\n",
    "          epochs=epochs,\n",
    "          batch_size=batches,\n",
    "          shuffle=False,\n",
    "          validation_data=(X_validate_std, y_validate),\n",
    "          callbacks=[EarlyStopping(monitor='val_loss', min_delta=0.001, patience=3, verbose=1, mode='auto'),\n",
    "          ReduceLROnPlateau(monitor='val_loss', factor=0.2, patience=1, min_lr=0.0001, verbose=1, mode='auto')])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import datetime\n",
    "model_filename = 'social-lstm_trained_model'+datetime.datetime.now().strftime(\"%Y%m%d%H%M\")+'.h5'\n",
    "model_LSTM.save(model_filename)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load previous trained model if required."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "model_LSTM = load_model('social-lstm_trained_model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## 9. Test on Testing Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "batches = 1\n",
    "y_result = model_LSTM.predict(X_test_std, batch_size=batches)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 10. Results: \n",
    "### The Probability of djia increase in next day is around 52%, Pr. of next 30 days increase is 85%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.5147661 ,  0.84850389],\n",
       "       [ 0.51476365,  0.8485049 ],\n",
       "       [ 0.51476359,  0.84850371],\n",
       "       [ 0.51476401,  0.84850472],\n",
       "       [ 0.51476401,  0.84850454],\n",
       "       [ 0.51476395,  0.84850472],\n",
       "       [ 0.51476467,  0.84850436],\n",
       "       [ 0.51476586,  0.84850407],\n",
       "       [ 0.51476383,  0.84850454],\n",
       "       [ 0.51476443,  0.84850407],\n",
       "       [ 0.51476437,  0.84850436],\n",
       "       [ 0.51476467,  0.84850436],\n",
       "       [ 0.51476449,  0.8485043 ],\n",
       "       [ 0.51476455,  0.84850436],\n",
       "       [ 0.51476383,  0.84850353],\n",
       "       [ 0.51476449,  0.8485043 ],\n",
       "       [ 0.51476729,  0.84850514],\n",
       "       [ 0.51476479,  0.84850454],\n",
       "       [ 0.5147649 ,  0.84850436],\n",
       "       [ 0.51476449,  0.84850436],\n",
       "       [ 0.5147664 ,  0.84850454],\n",
       "       [ 0.51476997,  0.84850311],\n",
       "       [ 0.51476544,  0.84850413]], dtype=float32)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 11. Actual Trend of T+1, T+30 of each day"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['0', '1'],\n",
       "       ['0', '1'],\n",
       "       ['0', '1'],\n",
       "       ['0', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['0', '1'],\n",
       "       ['0', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['0', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1'],\n",
       "       ['1', '1']], \n",
       "      dtype='<U1')"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEWCAYAAACufwpNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMFExDMMwwsb/A8K10zLCtAhQAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ce11a9bda0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import datetime as dt\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import matplotlib.pylab as plb\n",
    "df = djia.tail(23)\n",
    "_df = df.reset_index(level=1, drop=True) #Remove company information\n",
    "_df['close'].plot(color='r',label='DJIA')\n",
    "_df['T+1'].plot(color='g',label='T+1')\n",
    "_df['T+30'].plot(color='b',label='T+30')\n",
    "plt.title('Trend of DJIA during Sep 18, 2017 ~ Oct 18, 2017')\n",
    "plt.xlabel('Date')\n",
    "plt.ylabel('DJIA index')\n",
    "plt.legend(['DJIA', 'T+1', 'T+30'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbAAAAEWCAYAAAAHC8LZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ce12367710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = X.reset_index(level=1, drop=True) # Remove company information\n",
    "df = df.groupby(['date']).sum()\n",
    "social_df = df.tail(23) # Get last 23 days\n",
    "\n",
    "social_df['fblikes'].plot(color='blue', label='fblikes')\n",
    "social_df['fbdislikes'].plot(color='black', label='fbdislikes')\n",
    "social_df['twittelikes'].plot(color='cyan', label='twitter likes')\n",
    "plt.title('Trend of Social media likes during Sep 18, 2017 ~ Oct 18, 2017')\n",
    "plt.xlabel('Date')\n",
    "plt.ylabel('likes/dislikes number')\n",
    "plt.legend(['fblikes', 'fbdislikes', 'twitter likes'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "T+1 report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          0       0.00      0.00      0.00         7\n",
      "          1       0.70      1.00      0.82        16\n",
      "\n",
      "avg / total       0.48      0.70      0.57        23\n",
      "\n",
      "T+30 report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          1       1.00      1.00      1.00        23\n",
      "\n",
      "avg / total       1.00      1.00      1.00        23\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "threshold = 0.5\n",
    "y_result = np.where(y_result >= threshold, 1, 0)\n",
    "\n",
    "t1r = list(map(int,y_result[:,0]))\n",
    "t30r = list(map(int,y_result[:,1]))\n",
    "t1t = list(map(int,y_test[:,0]))\n",
    "t30t = list(map(int,y_test[:,1]))\n",
    "\n",
    "print('T+1 report\\n %s'%(metrics.classification_report(t1t, t1r))) # True value, predict results\n",
    "print('T+30 report\\n %s'%(metrics.classification_report(t30t, t30r)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[ 0  7]\n",
      " [ 0 16]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUUAAAEmCAYAAAD1FIKpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3XmcHFW5//HPNwkhYEDQAJqFsKMB\nEUnCz8smCmKACOgVBUGJIAhuICIiei9uXLjqD/WqV4yAgCibCsqigAsiyBbCGtnCmg1IRPYlC8/9\n45wJlclMT3dP9VTPzPedV73SVV196unq7mdOnTp1ShGBmZklQ6oOwMysnTgpmpkVOCmamRU4KZqZ\nFTgpmpkVOCmamRW0fVKUtJqkSyQ9LenCXpRzgKQry4ytKpJ2lHRvC8otZV9b+5C0gaSQNCzP/17S\nQU2Us76k5yQNLT/K9lJaUpT0YUkz8o5bkHf+DiUU/QFgPeD1EbFvs4VExC8iYrcS4mmp/AXepNY6\nEfG3iNi8BZvvdl9LOjV/ts9JWixpSWH+941sRNJ0SfdKekXStB7WPTNv79k83SXpJEmvbfjd1Rfb\nw5JezNt6StLfJR0uaUhhna9KOqcwv9JnJmlaXv7BVsTZrIjYPSLO6mm9vB92Lbzu0YgYGRHLWhth\n9UpJipKOBr4H/BfpR7U+8L/A3iUUPx64LyKWllBWv9fxF79Fut3XEXF4/lGMJH3O53fMR8TuXcR5\ntaSdu9nO7cAngZl1xvWtiFgDWAf4GPB24DpJr6nz9Y16b97eeOBk4IvA6Q2WcRDwZP6/FEra/uiu\n34uIXk3Aa4HngH1rrLMqKWnOz9P3gFXzczsDc4HPA08AC4CP5ee+BiwGluRtHAJ8FTinUPYGQADD\n8vw04EHgWeAh4IDC8msLr9sOuBl4Ov+/XeG5q4FvANflcq4ERnXz3jriP7YQ/z7AHsB9pB/G8YX1\ntwWuB57K6/4QGJ6fuya/l+fz+/1QofwvAo8BP+9Yll+zcd7GNnl+NLAI2LmbeN+c399TwCxgr+72\ndY3Pc4XPoJt1ru4uhsI61wLTeljnTOCbnZatkffdpwv74M/AP/N7/wWwVn7uC8CvO73+B8D3utne\nw8CunZZtC7wCbNnV+8+f2SaF+fF5/X8HlgLr1Xh/0/L37Af5u3gPsEun/XhiXudFYBPSb+70vA/m\nAd8Ehub1hwLfyfvhQeBTrPj7uBr4eKH8Q4G7Sd/zfwDb5O/YK3l7z5G+2xt0Kmc08Lv83ZsNHNrp\n+3EBcHYudxYwqbe5pq+mMpLilPzBD6uxzteBG4B1SX/t/w58Iz+3c37914FVSMnkBWDtbr6AneeX\nf1jAa4BngM3zc28Etih8+a7Nj18H/Av4SH7d/nn+9YUvzgPAZsBqef7kbt5bR/z/meM/FFgI/JL0\n490CeAnYKK8/kVTTGZZjvxs4qsYPrKP8/yb9cVmNQlLs9MVeHbgC+E43sa6Sv8DHA8OBd+Uv7eZd\n7dsan2eP69HCpJiXn02qrUJKFO/O+2cd0h+X7xW+A8/zapIcRvrjNbGb7T1Mp6SYlz8KHNHNd7Dz\nZ/YfwE358Z3A0TXe37T8+X4ufz4fIiXH1xX246P5ezQsr3Mx8BPS931d4CbgE3n9w0mJdRzpe/4X\nukmKwL6kpDoZUN6P47vaD6ycFP9KOhocAWxN+s7vUtg/L5F+y0OBk4Abeptr+moqoyr+emBR1D68\nPQD4ekQ8ERELSbWSjxSeX5KfXxIRl5P+OjXbZvYKsKWk1SJiQUTM6mKdPYH7I+LnEbE0Is4lfZHe\nW1jnZxFxX0S8SPqrt3WNbS4BToyIJcB5wCjg+xHxbN7+LGArgIi4JSJuyNt9mPTlfkcd7+mEiHg5\nx7OCiPgpcD9wIykJfLmbct4OjCQl+MUR8WfgUtIfhf5mPulHT0TMjoir8v5ZCJxC3qcRsYCUJDva\nSKeQvq+3NLu9OnyU9EeR/H9Ph9BPkJL4kog4H7iX9B3tcGZEzMq/sdcBu5P+kD4fEU8A3wX2y+t+\nMJc1JyKeJCWk7nyc1DRxcySzI+KRnt6cpHHADsAXI+KliLgNOI0Vf9PXRsTlkdogfw68tady20UZ\nSfGfwKge2rpGA8Wd/UhetryMTkn1BdKPtyER8TzpL+3hwAJJl0l6Ux3xdMQ0pjD/WAPx/DNebYDu\nSFqPF55/seP1kjaTdKmkxyQ9Q2qfG1WjbICFEfFSD+v8FNgS+EFEvNzNOqOBORHxSmFZ5/fdtHxi\n4ilJT5F+NJcWlh1XxjYKxpAO3ZC0rqTzJM3L+/QcVtynZwEH5scHkn6kTW+vFknbAxuS/jhCSopv\nkVTrj+q8yFWsrPPvY07h8XhSbXFBYV//hFRjJL+uuH6tJDeOdETUqNHAkxHxbKft1Pr9jGhxe3hp\nykiK15OqyvvUWGc+6cPssH5e1oznSYeJHd5QfDIiroiId5NqTPeQkkVP8XTENK/JmBrxY1Jcm0bE\nmqRDWfXwmppDGUkaSWqnPR34qqTuajTzgXGdGutLe98RsVbHRDo0nlpYdnIZ24Dl73dX4G950Umk\nfbRV3qcHsuI+vRjYStKWwFRSm2Mj25tM+sFfW8fqB+Vt3ybpMVLtHVLtsTtjJBXj7fz7KH7+c4CX\nSW3cHft2zYjYIj+/gJTsimV1Zw6pPbYrtb5z84HXSVqj03b64vfTcr1OihHxNKk97UeS9pG0uqRV\nJO0u6Vt5tXOBr0haR9KovP453ZXZg9uAnXK/qdcCX+p4QtJ6kvbKZyVfJh2Gd9WF4HJgs9yNaJik\nDwETSIeSrbYGqd3zuVyLPaLT848DGzVY5veBWyLi48BlwKndrHcj6Y/Ksfkz2pnUZHBeN+u3hKTh\nkkaQkscqkkbUc1ZV0qqSJpKS3L+An+Wn1iB91k9JGkM6ubJcrmX/ilRruykiHq0zzjUlTSXtn3Mi\n4s4e1h9BOnw9jNTc0jF9BjigRk1pXeCz+TPZl3Qy7PKuVszNAVcC/z/HN0TSxpI6mmAuyGWNlbQ2\nUKuGfhpwjKSJ+cz2JpI6Kgvdfg8jYg7pvMBJ+bPbinQStKE/Nu2qlNP7EXEKcDTwFVKD6xzg06Qv\nL6SzYzOAO0gNzzPzsma2dRVwfi7rFlZMZENIZ7Hnkw513kHq+tG5jH+SagyfJx3+H0uq1SxqJqYG\nHQN8mHSC46ek91L0VeCsfGjUYx83SXuT2skOz4uOBraRdEDndSNiMbAXqU1qEamh/KMRcU9zb6Vp\nV5KaFLYDpufHO9VY/1hJz5I+07NJn/t2ubkEUhv1NqQTFJcBv+mijLOAt1DfofMleXtzSO2zp5C6\nAhV1VZPaJ7+XsyPisY6JVIMfSvqcunIjsCnpMzkR+ED+jnbno6QTZf8g/XH4FenICNJ36gpSt6eZ\ndL0v0huIuDBv75ek7+PFvNpuehKpIvOUpGO6ePn+pJMv84GLSG3eV9WIud/Qik0ZZgOTpPVJzRZv\niIhnelnWKcCQiDiqhLimkc4Gl3Ghg5XAHUFtwMuH5kcD55WQENcC3kM68rEByEnRBrTcvvwMqR/j\nCb0sayrpbO2NpLY7ayOSzpD0hKS7Oi3/jNJlpbMK5zm6L8eHz2Y2EEjaiXTC7eyI2DIveyepXXjP\niHhZ0rq5b2e3XFM0swEhIq5h5b6kR5AuVng5r1MzIUK6bKhtjBo1KsaP36DqMKwkjz3bXR9y62+e\nenwezz/9ZE/9aRsydM3xEUtXukCrW/HiwlmkPtEdpkfE9B5ethmwo6QT82uPiYiba72grZLi+PEb\ncN2Nbr8eKL5z9eyqQ7CS/PiT7yu9zFj6IqtuXv/Iai/d9qOXImJSg5sZBqxNusR1MnCBpI2iRrth\nWyVFMxtMBK0fCW0u8JucBG+S9ArpEtCF3b3AbYpmVg0BUv1Tcy4mjQaFpM1Ind5rXqThmqKZVafE\nmqKkc0nD6o2SNJfUBesM4IzcTWcxcFCtQ2dwUjSzygiGlHfLl4jobgi8A7tZ3iUnRTOrTvOHxS3j\npGhm1RB9caKlYU6KZlaRXp1AaRknRTOrjmuKZmYFrimamXXok87bDXNSNLNqdHTebjNOimZWHdcU\nzcw6CIaW13m7LE6KZlYN91M0M+vEbYpmZh189tnMbEWuKZqZFbimaGaW9W7w2JZxUjSz6rimaGZW\n4JqimVkHn302M3uVKPV2BGVxUjSzirRnTbH9IjKzwaPEW5xKOkPSE/nOfZ2fO0ZSSBrVUzlOimZW\nHQ2pf+rZmcCUlTYhjQPeDTxaTyFOimZWnRJrihFxDfBkF099FzgWqHm/5w5uUzSzaqjhNsVRkmYU\n5qdHxPTam9BewLyIuF11dv9xUjSz6jTWT3FRREyqv2itDnwZ2K2RjTgpmlll6q29NWljYEOgo5Y4\nFpgpaduIeKy7Fzkpmlkl0i1aWpcUI+JOYN3l25MeBiZFxKJar/OJFjOrhoSG1D/1XJzOBa4HNpc0\nV9IhzYTlmqKZVabMmmJE7N/D8xvUU46ToplVpsVtik1xUjSzyjgpmpl1UJ7ajJOimVVCyDVFM7Mi\nJ0UzswInRTOzAidFM7MOPtFiZvYqIYYMab+L6pwUzawyPnw2Mytqv5zopGhmFZFrimZmK3BSNDMr\ncFI0M8t8mZ+ZWWftlxOdFPvSlVf8gWOOPpJly5Yx7eCP84Vjj6s6JGvSwjkPcsE3j1w+/6/H5vCu\ng45ku/d/rMKo+hmfaBncli1bxlGf/RSX/f4qxowdyw5vn8zUqXvx5gkTqg7NmrDOuI341E8uAeCV\nZcv49v47MGH7hm4aZ7RnUmy/7uQD1M033cTGG2/ChhttxPDhw9n3Q/tx6SW/rTosK8GDt/6d171x\nfdZab0zVofQ7Zd6jpSxOin1k/vx5jB07bvn8mDFjmTdvXoURWVnuvPoy3vLOqVWH0S9JqnvqKy1N\nipKmSLpX0mxJg7oBLSJWWtaOhw7WmKVLFnPP9X9my3fsXnUo/U4jCbGe34qkMyQ9IemuwrJvS7pH\n0h2SLpK0Vk/ltCwpShoK/AjYHZgA7C9p0DagjRkzlrlz5yyfnzdvLqNHj64wIivD/Tdfwxs3mcDI\ntUdVHUq/VHJN8UxgSqdlVwFbRsRWwH3Al3oqpJU1xW2B2RHxYEQsBs4D9m7h9trapMmTmT37fh5+\n6CEWL17Mheefx55T96o6LOulO/5yKVv50LlpZSbFiLgGeLLTsisjYmmevQEY21M5rUyKY4A5hfm5\nedkKJB0maYakGQsXLWxhONUaNmwY3/3+D3nvnu9h67e8mX/f94NM2GKLqsOyXlj80os8cMt1TNjx\nPVWH0n+pgQlGdeSKPB3W4NYOBn7f00qt7JLTVWpfqWEtIqYD0wEmTpy0csPbADJl9z2YsvseVYdh\nJRk+YjWO/83NVYfRrzXYrr4oIiY1uZ0vA0uBX/S0biuT4lxgXGF+LDC/hdszs/6kjzpvSzoImArs\nEl2d8eyklUnxZmBTSRsC84D9gA+3cHtm1o8IaHVOlDQF+CLwjoh4oZ7XtCwpRsRSSZ8GrgCGAmdE\nxKxWbc/M+hsxpMRO2ZLOBXYmtT3OBU4gnW1eFbgq10pviIjDa5XT0sv8IuJy4PJWbsPM+q8yD58j\nYv8uFp/eaDm+9tnMqqHWHz43w0nRzCohKPXwuSxOimZWGdcUzcwK2vH6fydFM6uG2xTNzF6V+im2\nX1Z0UjSzivjGVWZmK2jDnOikaGYVkbvkmJkt5zZFM7NO2jAnOimaWXVcUzQzK2jDnOikaGYV6aNB\nZhvlpGhmleiLQWab4aRoZhVx520zsxW0YU50UjSzirjztpnZq9x528ysk3ZMikOqDsDMBi+p/qnn\nsnSGpCck3VVY9jpJV0m6P/+/dk/lOCmaWWUk1T3V4UxgSqdlxwF/iohNgT/l+ZqcFM2sGg3UEuvJ\niRFxDfBkp8V7A2flx2cB+/RUjtsUzawSaryf4ihJMwrz0yNieg+vWS8iFgBExAJJ6/a0ESdFM6tM\ng+dZFkXEpBaFspyToplVZkjrzz4/LumNuZb4RuCJHmNqdURmZt0ps02xG78DDsqPDwJ+29MLXFM0\ns0pIMLTEK1oknQvsTGp7nAucAJwMXCDpEOBRYN+eynFSNLPKlNl5OyL27+apXRopp9ukKGnNHgJ4\nppENmZl11oYXtNSsKc4CgnSJYoeO+QDWb2FcZjbAidQtp910mxQjYlxfBmJmg08bDpJT39lnSftJ\nOj4/HitpYmvDMrMBr4FL/Ppy4Igek6KkHwLvBD6SF70AnNrKoMxscOiDLjkNq+fs83YRsY2kWwEi\n4klJw1scl5kNcKJPOm83rJ6kuETSENLJFSS9HnilpVGZ2aDQhjmxrjbFHwG/BtaR9DXgWuC/WxqV\nmQ0K7dim2GNNMSLOlnQLsGtetG9E3FXrNWZmPSn7ipay1HtFy1BgCekQ2tdLm1kp2i8l1nf2+cvA\nucBoYCzwS0lfanVgZjbw9cvDZ+BAYGJEvAAg6UTgFuCkVgZmZgNbOvtcdRQrqycpPtJpvWHAg60J\nx8wGjT6uAdar1oAQ3yW1Ib4AzJJ0RZ7fjXQG2sysV9owJ9asKXacYZ4FXFZYfkPrwjGzwaRf1RQj\n4vS+DMTMBpd+26YoaWPgRGACMKJjeURs1sK4zGwQaMeaYj19Ds8EfkZK7LsDFwDntTAmMxsEJBgq\n1T31lXqS4uoRcQVARDwQEV8hjZpjZtYr/XWUnJeV6rgPSDocmAf0eENpM7Oe9NfD588BI4HPAtsD\nhwIHtzIoMxscyqwpSvqcpFmS7pJ0rqQRPb9qZfUMCHFjfvgsrw40a2bWK0KljacoaQyp4jYhIl6U\ndAGwH+mcSENqdd6+iDyGYlci4v2NbszMbLny2wqHAatJWgKsDsxvtpDu/LCZAs06nPiF71UdgpXk\n5bmPt6TcBtsUR0maUZifHhHTASJinqTvkG54/yJwZURc2UxMtTpv/6mZAs3M6tXgOISLImJSV09I\nWhvYG9gQeAq4UNKBEXFOi2MyMyuHKHXosF2BhyJiYUQsAX4DbNdMXPUOMmtmVroSL/N7FHi7pNVJ\nh8+7ADNqv6RrdSdFSatGxMvNbMTMrLMyb0cQETdK+hUwE1gK3ApMb6asekbe3lbSncD9ef6tkn7Q\nzMbMzIqGqP6pJxFxQkS8KSK2jIiPNFuJq6dN8X+AqcA/84Zvx5f5mVkJ+utlfkMi4pFODZ3LWhSP\nmQ0Saeiw9rvMr56kOEfStkBIGgp8BrivtWGZ2WDQjt1f6kmKR5AOodcHHgf+mJeZmfVKG1YU67r2\n+QnSNYRmZqWRyrv2uUz1jLz9U7q4BjoiDmtJRGY2aLRhTqzr8PmPhccjgPcBc1oTjpkNJv3yHi0R\ncX5xXtLPgataFpGZDQqivM7bZWrmMr8NgfFlB2Jmg0ydnbL7Wj1tiv/i1TbFIcCTwHGtDMrMBgfR\nflmxZlLM92Z5K+m+LACvRES3A8+amdWrXe/7XLPvZE6AF0XEsjw5IZpZacq89rm0mOpY5yZJ27Q8\nEjMbdEocT7E0te7RMiwilgI7AIdKegB4nlTrjYhwojSzprXr4XOtNsWbgG2AffooFjMbTPp49Jt6\n1UqKAoiIB/ooFjMbZPrbZX7rSDq6uycj4pQWxGNmg0R/PHweCoyENuxIZGYDgBjaz2qKCyLi630W\niZkNKuluflVHsbIe2xTNzFqiTS/zq9VPcZc+i8LMBqUheUzFeqaeSFpL0q8k3SPpbkn/1kxM3dYU\nI+LJZgo0M6tHCw6fvw/8ISI+IGk4sHozhTQzSo6ZWSnK6pIjaU1gJ2AaQEQsBhY3FVMpEZmZNaHB\nW5yOkjSjMBVH/98IWAj8TNKtkk6T9JpmYnJSNLNKiJSA6p2ARRExqTBNLxQ3jHQF3o8j4m2kS5Kb\nGuLQSdHMqqFSB4SYC8yNiBvz/K9ISbJhTopmVhk1MNUSEY+R7lG/eV60C/CPZmLyiRYzq4Sg7Cta\nPgP8Ip95fhD4WDOFOCmaWWXKzIkRcRswqbflOCmaWUX6dvDYejkpmlklOs4+txsnRTOrjGuKZmYF\n7ZcSnRTNrCpyTdHMbDm3KZqZdeKaoplZQTsOMuukaGaVSIfP7ZcVnRTNrDJtePTspGhmVRFyTdHM\n7FWuKZqZZW5TNDMrkmuKZmYrcFI0MytoxxMt7XiVzYB15RV/YKstNmeLN23Ct791ctXhWINOPeEA\nHvnTScy48PgVlh+x3zu4/aL/4JZffZkTj9y7ouj6H5E6b9c79RXXFPvIsmXLOOqzn+Ky31/FmLFj\n2eHtk5k6dS/ePGFC1aFZnX5+yQ2cev5fOe0bH12+bKdJmzJ157cw+YMnsXjJUtZZe2SFEfY/Zd33\nuUyuKfaRm2+6iY033oQNN9qI4cOHs++H9uPSS35bdVjWgOtmPsCTT7+wwrLD9t2R7/zsKhYvWQrA\nwn89V0Vo/ZYa+NdXnBT7yPz58xg7dtzy+TFjxjJv3rwKI7IybDJ+XbZ/28Zcc/YxXHnakUycsH7V\nIfUb7Xr43LKkKOkMSU9IuqtV2+hPImKlZe04Qog1ZtjQIay95urs9NHvcPx3L+acbx1cdUj9SCP1\nxPp+K5KGSrpV0qXNRtXKmuKZwJQWlt+vjBkzlrlz5yyfnzdvLqNHj64wIivDvMef4uI/3Q7AjFmP\n8MorwSi3K9Yn91Osd6rTkcDdvQmrZUkxIq4BnmxV+f3NpMmTmT37fh5+6CEWL17Mheefx55T96o6\nLOulS66+g5233QyATdZfl+GrDGOR2xXr1t2N77uaeixLGgvsCZzWm5gqP/ss6TDgMIBx6w/c9phh\nw4bx3e//kPfu+R6WLVvGQdMOZsIWW1QdljXgrJOmsePETRm11khm/+EbfOPUyznr4uv5yVcPYMaF\nx7N4yTI+/p8/rzrMfiO1KTbUhDRK0ozC/PSImF6Y/x5wLLBGb+KqPCnmNzUdYOLESSs3vA0gU3bf\ngym771F1GNakg750ZpfLD/7K2X0byADSYKv6oojo8mb3kqYCT0TELZJ27k1MlSdFMxvEyjvXuD2w\nl6Q9gBHAmpLOiYgDGy3IXXLMrDJDpLqnWiLiSxExNiI2APYD/txMQoTWdsk5F7ge2FzSXEmHtGpb\nZtY/lXmipSwtO3yOiP1bVbaZDRAtyHYRcTVwdbOvd5uimVUi1QDb7wIGJ0Uzq4YHmTUzW1Eb5kQn\nRTOrUBtmRSdFM6uIb3FqZrYCtymamWV93f+wXk6KZlaZdhxT1EnRzCrThjnRSdHMqtOGOdFJ0cwq\n0qaNik6KZlYZd8kxM8uE2xTNzFbQhjnRSdHMKtSGWdFJ0cwq4zZFM7OCIe2XE50UzaxCTopmZolH\n3jYzK2rTkbd9i1Mzq0xZd/OTNE7SXyTdLWmWpCObjck1RTOrTnk1xaXA5yNipqQ1gFskXRUR/2i0\nICdFM6tIeSNvR8QCYEF+/Kyku4ExgJOimfUfDbYpjpI0ozA/PSKmr1ymNgDeBtzYTExOimZWiSYG\nyVkUEZNqlimNBH4NHBURzzQTl5OimVWnxLPPklYhJcRfRMRvmi3HSdHMKjOkpD45Svc1OB24OyJO\n6VVMpURkZtaEsrrkANsDHwHeJem2PO3RTEyuKZpZNUrsvB0R11LSwbiToplVqP0uaXFSNLNKeORt\nM7NO2jAnOimaWXVcUzQzK/DQYWZmRe2XE50Uzaw6bZgTnRTNrBpSeVe0lMlJ0cyq03450UnRzKrT\nhjnRSdHMqtOGR89OimZWlfJG3i6Tk6KZVaJdL/Pz0GFmZgWuKZpZZdqxpuikaGaVcZuimVmWOm9X\nHcXKnBTNrDpOimZmr/Lhs5lZQTueaHGXHDOrTIl380PSFEn3Spot6bhmY3JSNLPqlJQVJQ0FfgTs\nDkwA9pc0oZmQnBTNrDJq4F8PtgVmR8SDEbEYOA/Yu5mY2qpNcebMWxattooeqTqOPjAKWFR1EFaK\nwfJZji+7wFtn3nLF6sM1qoGXjJA0ozA/PSKm58djgDmF5+YC/6+ZuNoqKUbEOlXH0BckzYiISVXH\nYb3nz7J5ETGlxOK6qkpGMwX58NnMBoK5wLjC/FhgfjMFOSma2UBwM7CppA0lDQf2A37XTEFtdfg8\niEzveRXrJ/xZtoGIWCrp08AVwFDgjIiY1UxZimjqsNvMbEDy4bOZWYGToplZgZOimVmBk2IfkLS5\npH+TtEq+HMn6OX+OA5dPtLSYpPcD/wXMy9MM4MyIeKbSwKwpkjaLiPvy46ERsazqmKxcrim2kKRV\ngA8Bh0TELsBvSR1Mj5W0ZqXBWcMkTQVuk/RLgIhY5hrjwOOk2HprApvmxxcBlwLDgQ9L7TianHVF\n0muATwNHAYslnQNOjAORk2ILRcQS4BTg/ZJ2jIhXgGuB24AdKg3OGhIRzwMHA78EjiENTrA8MVYZ\nm5XLSbH1/gZcCXxE0k4RsSwifgmMBt5abWjWiIiYHxHPRcQi4BPAah2JUdI2kt5UbYRWBl/m12IR\n8ZKkX5BG7PhS/uG8DKwHLKg0OGtaRPxT0ieAb0u6h3Rp2TsrDstK4KTYByLiX5J+CvyDVMN4CTgw\nIh6vNjLrjYhYJOkO0mjP746IuVXHZL3nLjl9LDfKR25ftH5M0trABcDnI+KOquOxcjgpmvWCpBER\n8VLVcVh5nBTNzAp89tnMrMBJ0cyswEnRzKzASdHMrMBJcYCQtEzSbZLuknShpNV7UdbOki7Nj/eS\ndFyNddeS9MkmtvFVScfUu7zTOmdK+kAD29pA0l2NxmiDk5PiwPFiRGwdEVsCi4HDi08qafjzjojf\nRcTJNVZZC2g4KZq1KyfFgelvwCa5hnS3pP8FZgLjJO0m6XpJM3ONciSApCmS7pF0LfD+joIkTZP0\nw/x4PUkXSbo9T9sBJwMb51rqt/N6X5B0s6Q7JH2tUNaXJd0r6Y/A5j29CUmH5nJul/TrTrXfXSX9\nTdJ9eUgvJA2V9O3Ctj/R2x1pg4+T4gAjaRjpsrM786LNgbMj4m3A88BXgF0jYhvSgLdHSxoB/BR4\nL7Aj8IZuiv8f4K8R8VZgG2AWcBzwQK6lfkHSbqSh0rYFtgYmStpJ0kTSvXjfRkq6k+t4O7+JiMl5\ne3cDhxSe2wB4B7AncGp+D4dVu2HFAAAByklEQVQAT0fE5Fz+oZI2rGM7Zsv52ueBYzVJt+XHfwNO\nJ43E80hE3JCXvx2YAFyXh3IcDlwPvAl4KCLuB8gjvxzWxTbeBXwUlg+X9XS+1K1otzzdmudHkpLk\nGsBFEfFC3kY9NyrfUtI3SYfoI0n39O1wQb5U8n5JD+b3sBuwVaG98bV52/fVsS0zwElxIHkxIrYu\nLsiJ7/niIuCqiNi/03pbk0bxKYOAkyLiJ522cVQT2zgT2Ccibpc0Ddi58FznsiJv+zMRUUyeSNqg\nwe3aIObD58HlBmB7SZsASFpd0mbAPcCGkjbO6+3fzev/BByRXzs031LhWVItsMMVwMGFtsoxktYF\nrgHeJ2k1SWuQDtV7sgawIN/W4YBOz+0raUiOeSPg3rztI/L6SNosj5htVjfXFAeRiFiYa1znSlo1\nL/5KRNwn6TDgMkmLSKODb9lFEUcC0yUdAiwDjoiI6yVdl7u8/D63K74ZuD7XVJ8jDZM2U9L5pFHH\nHyEd4vfkP4Ab8/p3smLyvRf4K2lcysPzuJWnkdoaZyptfCGwT317xyzxgBBmZgU+fDYzK3BSNDMr\ncFI0MytwUjQzK3BSNDMrcFI0MytwUjQzK/g/JsXHngJzkGAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ce1241c940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import itertools\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "    print(cm)\n",
    "\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n",
    "\n",
    "\n",
    "\n",
    "class_names = ['0','1']\n",
    "# Compute confusion matrix\n",
    "cnf_matrix = confusion_matrix(t1t, t1r)\n",
    "np.set_printoptions(precision=2)\n",
    "\n",
    "# Plot non-normalized confusion matrix\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cnf_matrix, classes=class_names,\n",
    "                      title='Confusion matrix of T+1 Day DJIA prediction')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}